Guys, please don't feed it. Too transparent, absolutely uncreative, save it for a more skilled one.
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Silent Angel Bonn N8 Audio Grade Ethernet Switch
- Thread starter amirm
- Start date
Hi
I shudder at the poor person who, having conducted a search on google about " ASR +Silent Angel Bonn N8 Audiophile switch" would find your "review" ...
Trying to decide if you're sincere or a Master Troll...
...
You're the latter.
Disagree. Not even basic competency, much less mastery.
- Joined
- Feb 28, 2016
- Messages
- 1,279
- Likes
- 1,180
Here is my $8000 to your $1000 offer. You hit 13 of 15 changes $8000 is yours. Loser pays travel expenses. Of course you won't know when a change has been made or even see the switch...
We do this in your audio rig from your USB cable on back.
Last edited:
As mentioned TCP/IP and Ethernet would not work if the equipment we are using everyday needed special snake oil.
Clock upgrade do not make sense what so ever to me on these switch devices.
On the other-hand can't we have audio equipment (not always built to the highest EMI/RFI standards) being affected by stray current from the network equipment (Cheap switches with poor SMPS wall warts)?
I did a small experiment at home:
I have a network with ISP router, Synology router, and going to my hifi setup a D-Link DGS-105. The DGS 105 is plugged to a SOTM SMS 200 Ultra which in turn is plugged to the dac.
The DGS-105 is powered by its 5v 1A SMPS wall wart. I made a quick & dirty Linear PSU with a 9v 1.2A transformer and a 7805.
To my surprise, there is a clear sound difference between the set up where the D-Link DGS-105 is powered by its standard SMPS and the set up where the D-Link DGS-105 is powered by the quick and dirty linear psu. The sound appear softer in the latter.
My wife who was in the kitchen could hear the difference as well.
Unfortunately I do not have an Audioprecision nor would I be able to use it.
Besides electrical "pollution" going up the chain I do not see what could be the cause.
Any idea (No I am not on drugs)?
Clock upgrade do not make sense what so ever to me on these switch devices.
On the other-hand can't we have audio equipment (not always built to the highest EMI/RFI standards) being affected by stray current from the network equipment (Cheap switches with poor SMPS wall warts)?
I did a small experiment at home:
I have a network with ISP router, Synology router, and going to my hifi setup a D-Link DGS-105. The DGS 105 is plugged to a SOTM SMS 200 Ultra which in turn is plugged to the dac.
The DGS-105 is powered by its 5v 1A SMPS wall wart. I made a quick & dirty Linear PSU with a 9v 1.2A transformer and a 7805.
To my surprise, there is a clear sound difference between the set up where the D-Link DGS-105 is powered by its standard SMPS and the set up where the D-Link DGS-105 is powered by the quick and dirty linear psu. The sound appear softer in the latter.
My wife who was in the kitchen could hear the difference as well.
Unfortunately I do not have an Audioprecision nor would I be able to use it.
Besides electrical "pollution" going up the chain I do not see what could be the cause.
Any idea (No I am not on drugs)?
Last edited:
You are imaging the difference. It's that simple.
Can't we get more original than this?
Maybe the difference, if real, isn't the smps polluting the dlink, but polluting other devices plugged into the mains.
That's the variable you're not testing for. Plug the dlink smps into the mains and use it to power any device not connected to the hifi and see if the difference comes back.
That's the variable you're not testing for. Plug the dlink smps into the mains and use it to power any device not connected to the hifi and see if the difference comes back.
- Joined
- Oct 11, 2018
- Messages
- 3,739
- Likes
- 6,449
Nice play. Some of the best humor is both subtle, and creeps up on you unexpectedly.
And the fact that your wife is actually in the kitchen has to be a plus. If you ask me.
Its a well known fact women have much better hearing than men.
For example , I have never heard my wife fart .
But, even if I am at the bottom of the garden and she is in the kitchen , she will hear me drop one.
What more proof do you need?
For example , I have never heard my wife fart .
But, even if I am at the bottom of the garden and she is in the kitchen , she will hear me drop one.
What more proof do you need?
Hi,
this thread is a bit long, I admit I just read the 1st post from Amir & then jumped to this last page...
so maybe my remark is in the posts behind...
Amir made some measurements, & concluded : no difference between "audiophile" & basic switches.
These measurements were done the right way ? no.
Studying this type of device by looking at the audio bandwidth only is a mistake (I would say... it's a closer to a "joke" than a "mistake").
These devices work at much higher frequencies than 20kHz. They deal with digital stream (real-time). Thus, putting probe on them like on a valve amp is a non-sense.
You have to look at higher frequencies (>MHz).
Amir, you stand you have top-notch measurement devices. I guess you have an oscillo reaching 200MHz or so, and a spectrum analyzer reaching 2GHz min : please have a look in these higher frequencies to give a conclusion with correct added-value.
2nd remark :
- often seen, but quite annoying after a while : all these posts that do bashing about "audiophile" switch or simply the effect of a switch on SQ
- of course, each time, from people that never run tests of course... LOL
Rgds
this thread is a bit long, I admit I just read the 1st post from Amir & then jumped to this last page...
so maybe my remark is in the posts behind...
Amir made some measurements, & concluded : no difference between "audiophile" & basic switches.
These measurements were done the right way ? no.
Studying this type of device by looking at the audio bandwidth only is a mistake (I would say... it's a closer to a "joke" than a "mistake").
These devices work at much higher frequencies than 20kHz. They deal with digital stream (real-time). Thus, putting probe on them like on a valve amp is a non-sense.
You have to look at higher frequencies (>MHz).
Amir, you stand you have top-notch measurement devices. I guess you have an oscillo reaching 200MHz or so, and a spectrum analyzer reaching 2GHz min : please have a look in these higher frequencies to give a conclusion with correct added-value.
2nd remark :
- often seen, but quite annoying after a while : all these posts that do bashing about "audiophile" switch or simply the effect of a switch on SQ
- of course, each time, from people that never run tests of course... LOL
Rgds
By the way, I posted this => https://www.audiosciencereview.com/...irs-music-server-build-thread.415/post-949824
a while ago
it's in close relationship with this thread about switches
still waiting for answer... thanks in advance for reply etc...
Rgds
a while ago
it's in close relationship with this thread about switches
still waiting for answer... thanks in advance for reply etc...
Rgds
radix
Major Contributor
- Joined
- Aug 1, 2021
- Messages
- 1,397
- Likes
- 1,334
I think Amir's test is perfectly valid black-box testing of the Ethernet switch. Did the switch make any measurable difference in the delivered audio stream? No.
Amir was not testing any differences in the Ethernet frames. And I think that is pointless for audio for several reasons.
1000BASE-T uses 4D-PAM5 (4 dimensional pulse amplitude modulation with 5 bits/symbol) convolution coding of the bitstream at 125 MBd with a scrambler. The receiving unit recovers the clock frequency of the symbols from the training header. 125 MBd is 8 nano-second symbols. I don't know what the recovered clock accuracy is, but I expect it is in the pico-seconds for the length of a packet (about 12 micro-seconds). All Ethernet switches do this in their ASIC silicon. In the silicon's physical layer, it must receive all the bits, convolution decode them, de-scramble them, re-create the 802.2 Ethernet frame, and pass that up to the next layer. That next layer then checks the Ethernet CRC and likely does UDP/TCP checksum offload. Once that's done it can finally leave the ASIC silicon, and there's DMAs over the PCI bus to main memory. Symbol-level timing and bit-level timing will be entirely invisible to the upper-layer protocols unless its entirely broken.
Ethernet frame jitter should be entirely masked by receiver buffering. The re-construction of the audio signal uses timestamps inside the transport protocol, it does not depend on minimizing the Ethernet jitter. That is why streaming protocols work over the Internet where you can have orders of magnitude different jitters.
Pretty much all modern Ethernet switches and networks are collision-free and for most practical purposes contention-free for home networks, unless you are copying giant movies all to the same ports. They are collision-free because all links are full-duplex. There is never any possibility of more than 1 device transmitting on an uplink. They are contention-free because home networking traffic is usually so far below the switching capability of the hardware ASIC even in cheap switches there's no significant queuing (there only exception is when multiple transmitters are going to the same egress port, and even then they likely timeshare well at the Ethernet level).
Load contention is pretty much the only thing I think would disrupt a streaming protocol, and modern ethernet switches allow you to set priority values. If you know you have a messed up network load, use DSCP. Or switch to 10 Gbps uplinks on that messed up egress port.
I am not sure where 2 GHz came from. Are you saying he should look at the audio signal out to 2 GHz? And looking at an Ethernet link at 200 MHz will not help much, you need special analyzers to evaluation the quality of the eye diagrams.
What would be helpful is to describe a specific mechanism by which improving an Ethernet switch could correlate to improving an audio streaming protocol.
Marc
5 levels, 2 bits per symbol.
More handwaving nonsense. If the audio is the same the switch doesn't matter. Maybe it could with some shit dac on the end of it, but I've yet to hear of any ethernet equipped dac that measures differently fed from a different, switch/ router/ ethernet source.
Hi Marc,
thanks for the explanation about "how does it work", but it is not the issue here.
The issue is : SQ can be improved by using switch A instead of switch B (no matter is "audiophile, a basic Buffalo is better than some others etc...).
Then, the question is : why ? How can it be possible ? Which part of the process can affect SQ as such ?
The data transfer is fine of course, 100% accurate etc.
But it sounds different.
One explanation (test yourself & you'll see) is the power delivery network (PDN) on the PCBs (noise of regs / interaction between components creating micro-ripples etc). All stuffs related finaly to the transient responses on the PCB.
These transient states play a role into these diff in SQ, for instance by creating noise, that affects PLL for instance that affects the clocking/unclocking of the signal (clock is embedded into the data ; here, compared to i2s, no specific "clock" signal along "data" signal".
Then, if you stand that :
- "no SQ diff between switches"... that's another story.... the story that consider all users / tweakers of switches in thier audio gear are all jerks... Change the clock (oscillator or quartz) on any switch and you'll hear a clear improvement in SQ.
- "network devices are perfect", so "move along, no issue there". Why not, but that's not an decent explanation, isn't it ?
- Amir measurements were OK ? Cool ! But to be clearer, doing measurements on the audio bandwidth of a network device... that's as weird as using a stethoscope on the same swtich) LOL Of course, the transient states I told above don't occur in the audio bandwidth LOL !
Of course it is not a "network" issue, since everything reach the end of the tunnel perfectly. It's more a basic "electrical" issue.
Rgds
thanks for the explanation about "how does it work", but it is not the issue here.
The issue is : SQ can be improved by using switch A instead of switch B (no matter is "audiophile, a basic Buffalo is better than some others etc...).
Then, the question is : why ? How can it be possible ? Which part of the process can affect SQ as such ?
The data transfer is fine of course, 100% accurate etc.
But it sounds different.
One explanation (test yourself & you'll see) is the power delivery network (PDN) on the PCBs (noise of regs / interaction between components creating micro-ripples etc). All stuffs related finaly to the transient responses on the PCB.
These transient states play a role into these diff in SQ, for instance by creating noise, that affects PLL for instance that affects the clocking/unclocking of the signal (clock is embedded into the data ; here, compared to i2s, no specific "clock" signal along "data" signal".
Then, if you stand that :
- "no SQ diff between switches"... that's another story.... the story that consider all users / tweakers of switches in thier audio gear are all jerks... Change the clock (oscillator or quartz) on any switch and you'll hear a clear improvement in SQ.
- "network devices are perfect", so "move along, no issue there". Why not, but that's not an decent explanation, isn't it ?
- Amir measurements were OK ? Cool ! But to be clearer, doing measurements on the audio bandwidth of a network device... that's as weird as using a stethoscope on the same swtich
) LOL Of course, the transient states I told above don't occur in the audio bandwidth LOL !Of course it is not a "network" issue, since everything reach the end of the tunnel perfectly. It's more a basic "electrical" issue.
Rgds
OP
- Thread Starter
- #238
You want me to measure things you can't hear?
If your answer is yes, well I measured that as well:
That test goes to 100 kHz or five to ten times your hearing bandwidth and still no difference.
That there is not even a trace of any effect should tell you that you are chasing ghosts here if you think there is an audible difference.
OP
- Thread Starter
- #239
If you dreamed you were flying while dreaming, do you expect a physical test to show how you were able to fly while awake?
Your brain is subject to imagination. It can be trivially fooled. I can take two identical audio boxes and paint one red and the other black. If I tell you they are different things, you will perceive them as sounding different. You will think the black one sounds "darker." The only way to avoid this is to test you blind where you can't tell the two are different colors. When we do that in controlled testing, the observation that the sound is different disappears. Once there, it matches the engineering and measurements of the devices which shows there is no audible difference.
So look within for the answer to that question. That both by design and comprehensive measurements we know these switches make no audible difference. Therefore you need to learn why you perceive a difference that is not physically there. Once you solve that equation, you will forever be transforming your view of audio. You will save a ton of money not chasing ineffective fixes and tweaks for audio. Put that toward a better purpose like more music, great food, a vacation, etc. Don't waste it on non-existent things that are so easily explained.
OP
- Thread Starter
- #240
Not jerks. Just not knowledgeable. Huge number of people worldwide associate getting a "cold" with cold weather. Winter comes and a ton of people get cold so they think the two are equal and hence the name. Science of course says it is a virus and spreads more in winter due to people staying more indoor. It has little to do with whether being colder. The fact that these people go by lay intuition of their personal observation doesn't make them "jerks." It makes them human.
Now you can prove us wrong by changing said clock and have someone test you blind with and without said change a dozen times. If you are right, you should be able to tell 100% of times this is the case. But you won't be able to as such evidence has never been presented during all these years folks like you chase ineffective changes.
Remember, we all can and do experience what you say. We are human as well.
The difference between us is that we have awareness of mistakes our brain makes in drawing conclusions and you are not. Same as your doctor knowing more about medicine than you if you are not a doctor. Or a lawyer about the law. Don't be your own brain surgeon as the saying goes. Listen to people who have studied these things extensively. Don't be dismissive with lay logic and a few technical buzz words.Similar threads
